©istock.com/Tatiana Alex
Click here for this issue’s Table of Contents
Every Month, the Editor-in-Chief selects an Article of the Month from the current issue of BJUI. The abstract is reproduced below and you can click on the button to read the full article, which is freely available to all readers for at least 30 days from the time of this post.
In addition to the article itself, there is an accompanying editorial written by a prominent member of the urological community. This blog is intended to provoke comment and discussion and we invite you to use the comment tools at the bottom of each post to join the conversation.
If you only have time to read one article this month, it should be this one.
The European Organization for Research and Treatment of Cancer (EORTC) randomized trial 30904 reported that for solitary renal masses ≤5 cm, radical nephrectomy (RN) was associated with a higher overall survival (OS; primary endpoint): 81%, compared with 76% for nephron-sparing surgery (NSS) at a median follow-up of 9.3 years (P = 0.03). The difference in cancer-specific mortality, however, was not significant. For histologically proven RCC, and after exclusion of patients with positive surgical margins, NSS was associated with equivalent OS compared with RN [1]. It is noteworthy that the renal function outcomes of the two groups in the trial have been reanalysed, showing that renal function does not decline over time after RN, as was expected [2].
The EORTC 30904 trial had difficulty recruiting and randomizing patients, and was criticized for not meeting the accrual goal of 1300 patients. Additionally, the generalizability of the study findings to ‘real-world’ patients has been questioned. Despite the criticism, and more than 20 retrospective studies [3, 4] showing better OS and cancer-specific survival with NSS, this randomized clinical trial (RCT) remains the only available level 1 evidence on this subject. Notably, no study to date has formally examined the external validity [5] of the trial.
For any RCT to be externally valid, its supposedly randomly selected sample must be representative of the general population seen in clinical practice. In this context, we studied patients with localized RCC treated with NSS or RN within the National Cancer Database (NCDB), in an effort to test the external validity of the EORTC 30904. Our objective was not to compare survival outcomes between the two treatment arms, as this is beyond the scope of examining the external validity of an RCT, and such analysis is already available in literature. Instead, our aim was to ascertain if the trial patients were representative of contemporary patients with RCC in the USA, using the NCDB, which captures ~70% of all incident cancer diagnoses in the USA [6].
We identified patients who met the clinical and pathological inclusion criteria of the EORTC 30904 within the NCDB from 2004 to 2013: histologically confirmed RCC; tumour size ≤5 cm; clinically node-negative, non-metastatic disease; no positive surgical margins; and no pT3/4 disease. After exclusions, there were 90 844 assessable patients within the NCDB, of whom 41 588 (45.78%) underwent RN and 49256 (54.22%) underwent NSS. The demographic characteristics, namely, age, gender (percentage of men), presence of comorbidities (yes/no), histology (clear cell/non-clear cell), Fuhrman grade (1, 2, 3 or 4) and surgical approach (open/robotic/laparoscopic) were then compared with the patients enrolled in the EORTC 30904. The statistical significance of differences in categorical variables was tested using the chi-squared test. Unfortunately, the trial did not provide measures of variance (such as standard deviation, or interquartile range) for continuously coded variables; we were therefore unable to test for the statistical significance of differences in these variables. All analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA), with a P value <0.05 taken to indicate statistical significance.
The median age of the NCDB cohort was 60.0 years, compared with 62.0 years in the EORTC 30904. The median clinical tumour size in the NCDB was 30 mm, similar to the 30-mm tumour size observed in the trial. The percentage of men was 59.4% in the NCDB vs 65.8% in EORTC 30904 trial (P < 0.001). The NCDB cohort was healthier, with 70.03% patients having no comorbidity vs 62.8% in the trial (P < 0.001). The percentage of patients with clear-cell histology was 81.9% in the NCDB vs 62.9% in the trial (P < 0.001). The trial did not report data on race, while the NCDB had 15.6% non-white patients. Finally, the percentage of patients with high-grade disease (Fuhrman grade ≥3) was 21.1% in the NCDB vs 11.2% in the EORTC 30904 (P < 0.001; Table 1). Notably, in the EORTC 30904 trial, there was no central pathology review.
| Variable | EORTC 30904 trial | NCDB | P |
|---|---|---|---|
|
|||
| Study period | 1992–2003 | 2004–2013 | – |
| Number of patients clinically and pathologically eligible | 391 | 90 844 | – |
| Median (IQR) age, years | 62 (not provided) | 60 (51–69) | – |
| Median (IQR) clinical tumour size, mm | 30 (not provided) | 30 (21–40) | – |
| Men, % | 65.8 | 59.44 | <0.001 |
| Race | Not provided | Non-white 15.6% | – |
| Free of comorbid disease, % | 62.8 | 70.03 | <0.001 |
| Clear cell histology, % | 62.9 | 81.9 | <0.001 |
| Tumour grade, % | |||
| 1 | 22.30 | 18.08 | <0.001 |
| 2 | 66.60 | 60.78 | |
| 3 | 10.50 | 19.63 | |
| 4 | 0.70 | 1.51 | |
| Surgical approach (recorded in NCDB since 2010, n = 15 604), % | |||
| Open | 100 | 39.4 | <0.001 |
| Robotic | 0 | 34.6 | |
| Laparoscopic | 0 | 23.5 | |
Several important observations emerge from these results. First, age and tumour size were similar in the EORTC 30904 trial and the NCDB. These two variables are the most important determinants of mortality and stage of disease, respectively, which implies that the trial was able to recruit patients representative of those seen in ‘real-world’ clinical practice.
Second, there was a higher incidence of high-grade disease and clear-cell histology in the NCDB cohort compared with the EORTC 30904 trial. In other words, patients in the NCDB had more aggressive tumours as compared with patients in the trial. Arguably, such patients are better served with RN, which has a higher probability of completely eradicating the tumour. The survival benefit of RN observed in the trial might therefore be even more evident in clinical practice, where a higher proportion of patients harbour unpredictable aggressive disease.
Finally, the EORTC seems to have recruited patients with a higher comorbidity burden than is generally observed in clinical practice. The significance of this finding is controversial. On the one hand, it might be argued that the higher background mortality of the cohort could have masked the potential OS benefit of NSS by offering this treatment method to sicker patients with limited life expectancy [7]. On the other hand, preserving renal function might be even more important in sicker patients, who have the burden of other comorbidities [8].
The present study has some limitations. An inherent limitation of the NCDB is the lack of information on the performance status of patients. Second, the comparison was between two cohorts separated in time. The mode of treatment and thus, patient selection might have changed over time. The NCDB provides information about surgical approach starting in 2010, and indeed open surgery was performed in only 39.4% of the cases compared with 100% in the trial. More than 15% of patients in the NCDB had missing tumour grade compared with 4% in the trial; however, this proportion was equally distributed between patients undergoing RN and PN in the NCDB (data not shown). Despite the limitations, these findings are significant in the context of the recent debate on contemporary guidelines recommending NSS ‘wherever possible’ in patients with a normal contralateral kidney [9].
In conclusion, our results indicate that, although the EORTC 30904 cohort had somewhat different baseline characteristics than ‘real-world’ patients with small renal masses, none of these differences seem to have the potential to significantly alter the outcomes of the trial. The latter should therefore be considered generalizable to contemporary North American patients with renal masses ≤5 cm.
In this issue of the BJUI, Abdollah et al. [1] have for the first time tested the external validity of the only randomized clinical trial, 30904, run by the European Organization for the Research and Treatment of Cancer Genito-Urinary Group (EORTC GU) in the early 1990s, comparing cancer-specific survival and overall survival in patients with solitary renal masses of ≤5 cm and stage T1 and T2 in the TNM classification (in use at that time). The trial showed, as expected, that renal function was worse after radical nephrectomy (RN) and that the complication rate was higher after partial nephrectomy (PN) [2]. However, unexpectedly, overall survival after PN was not better than after RN [3], as was suggested or claimed in many non-randomized studies and also in a meta-analysis that included the EORTC 30904 trial as the only randomized clinical trial [4].
Despite a couple of limitations in the randomized trial, and it’s premature closure because of slow accrual, we performed a second analysis looking at the estimated GRF in the vast majority of the included patients and, most importantly, showed that kidney function did not progressively deteriorate after RN when the contralateral kidney was normal, and that only exceptionally did patients developed chronic kidney disease (CKD) necessitating dialysis [5].
Whilst it was anticipated that decreased kidney function should induce cardiovascular disease and increase cardiovascular death, this was separately investigated by Capitanio et al. [6] in a multicentre study where this suggestion was confirmed. However, looking at their Kaplan–Meier curves, it is clear that, although the negative impact on cardiovascular disease should become more and more obvious and accumulate over time, the split of the curves in favor of PN occurred very early after surgery. This indicates that the patients included in these non-randomized studies were different from the start, meaning that those selected for PN were ‘better’ patients who obviously had less cardiovascular disease and therefore had better cardiovascular outcomes. Another study confirmed that both PN and RN impact on cardiovascular disease [7], whilst another meta-analysis showed no difference for cardiovascular outcomes [8]. Obviously patients with preoperative CKD will benefit from nephron-sparing surgery [9], as well as those who have concomitant conditions, e.g. hypertension, diabetes, and a worse Charlson’s Comorbidity Index [10].
The authors, who tested the external validity of the EORTC 30904 trial in contemporary North American patients, need to be congratulated for the effort undertaken to show that the EORTC 30904 cohort was not significantly different from the National Cancer Database cohort in a manner that could influence the reported trial outcomes.
Every Week the Editor-in-Chief selects an Article of the Week from the current issue of BJUI. The abstract is reproduced below and you can click on the button to read the full article, which is freely available to all readers for at least 30 days from the time of this post.
In addition to the article itself, there is an accompanying editorial written by a prominent member of the urological community. This blog is intended to provoke comment and discussion and we invite you to use the comment tools at the bottom of each post to join the conversation.
Finally, the third post under the Article of the Week heading on the homepage will consist of additional material or media. This week we feature a video from Archie Fernando and Tim O’Brien, discussing their paper.
If you only have time to read one article this week, it should be this one.
To determine the scope and outcomes of nephron-sparing surgery (NSS), i.e. partial nephrectomy, across the UK and in so doing set a realistic benchmark and identify fresh contemporary challenges in NSS.
In 2012 reporting of outcomes of all types of nephrectomy became mandatory in the UK. In all, 148 surgeons in 86 centres prospectively entered data on 6 042 nephrectomies undertaken in 2012. This study is a retrospective analysis of the NSS procedures in the dataset.
A total of 1 044 NSS procedures were recorded and the median (range) surgical volume was 4 (1–39) per consultant and 8 (1–59) per centre. In all, 36 surgeons and 10 centres reported on only one NSS. The indications for NSS were: elective with a tumour of ≤4.5 cm in 59%, elective with a tumour of >4.5 cm in 10%, relative in 7%, imperative in 12%, Von Hippel–Lindau in 1%, and unknown in 11%. The median (range) tumour size was 3.4 (0.8–30) cm. The technique used was minimally invasive surgery in 42%, open in 58%, with conversions in 4%. The histology results were: malignant in 80%, benign in 18%, and unknown in 2%. In patients aged <40 years 36% (36/101) had benign histology vs 17% (151/874) of those aged ≥40 years (P < 0.01). In patients with tumours of <2.5 cm 29% (69/238) had benign histology vs 14% (57/410) with tumours of 2.5–4 cm vs 8% (16/194) with tumours of ≥4 cm (P = 0.02). In patients aged <40 years with of tumours of <2.5 cm 44% (15/34) were benign. The 30-day mortality was 0.1% (1/1 044). There were major complications (Clavien–Dindo grade of ≥IIIa) in 5% (53/1 044). There was an increased risk of complications after extended elective NSS of 19% (19/101) vs elective at 12% (76/621) (relative risk [RR] 1.54; P < 0.01). Margins were recorded in 68% (709/1 044) of the patients, with positive margins identified in 7% (51/709). Positive surgical margins after NSS for pathological T3 (pT3) tumours were found in 47.8% (11/23) vs 6.1% (32/523) for pT1a, tumours (RR 5.61; P < 0.01). In all, 14% (894/6 042) of the patients underwent surgery for T1a tumours: 55% (488/894) by NSS, 42% (377/894) by radical nephrectomy (RN), and in 3% (29/894) the procedure used was unknown. Major complications after occurred in 4.9% (24/488) of NSS vs 1.3% (5/377) of RN (P < 0.01). Limitations included poor reporting of renal function data and no data on tumour complexity.
In its first year, mandatory national reporting has provided several challenging contemporary insights into NSS.
The perceived wisdom that a small enhancing mass in the kidney represents a surgical lesion that automatically requires excision without the need for a preoperative biopsy has been challenged by Fernando et al. [1] in this issue of BJUI.
The authors are to be congratulated in bringing these data to publication to provoke debate on the treatment paradigm for small renal masses (SRMs) by reviewing nationally collected data on the main therapeutic surgical option: nephron-sparing surgery. As anyone who has attended a renal multidisciplinary meeting can testify, the predominant presentation of renal cancer is the incidentally detected SRM, often in elderly patients with significant comorbidity.
As the authors emphasize, these data are unique in representing a national picture encompassing both high- and low-volume centres, as opposed to the majority of the studies in the literature, which report data from high-volume tertiary referral centres.
Drawing conclusions from data requires a clear understanding of the source and quality. Most importantly, as these data only refer to patients undergoing nephron-sparing surgery, we need to be cautious about extrapolating to infer information on the management of SRMs in general.
For instance, a striking finding of the present study is the high incidence of benign lesions in the younger age groups. We have no knowledge of the numbers of patients with SRMs within the study period who had biopsy-proven benign disease and thus avoided surgery. It is probable that the true incidence of benign disease would be even higher if these cases had been recorded and included in the analysis.
An inherent difficulty with self-reported data is the issue of compliance, and this is clearly evident in the present study, with, for example, almost a third of cases missing data on surgical margin results. It would perhaps be helpful for future audits if the BAUS dataset had a clear definition of positive surgical margin in recognition of the surgical drift to enucleation rather than excision with a margin of renal parenchyma.
The variation in caseload between reporting centres raises important questions, as does the finding that two fifths of patients with T1a tumours underwent radical nephrectomies. As the authors concede, with the numbers involved and the absence of any measure of tumour complexity, it is difficult to draw firm conclusions; however, the study does highlight the need to examine this issue in future analyses and to consider including some form of renal scoring system in future audits.
Where do we go from here and what can we do with this information? First, we need to rethink our discussion with patients with SRMs. Can we justify performing major surgery with a one in 20 chance of a significant complication for a possible benign lesion without at least a pragmatic discussion of the role of renal biopsy with the patient? Indeed, one may argue, could it really be an ‘informed’ decision without it?
Second, we need to improve the quality of the data by encouraging robust data reporting, increasing the completion rate and considering adding data fields which will allow us to draw clearer conclusions on surgical margin and surgical outcome and volume relationships.
Third, we need to recognize that nephron-sparing surgery is only one component of the management of SRMs, which represents a major contemporary challenge in terms of health resources and, most importantly, in deciding the best treatment paradigm for our patients. If BAUS can carry out this audit, could we not extend this to all patients with SRMs, whether they have surgery, ablation or surveillance, and establish greater clarity on these treatment methods?
To determine the scope and outcomes of nephron-sparing surgery (NSS), i.e. partial nephrectomy, across the UK and in so doing set a realistic benchmark and identify fresh contemporary challenges in NSS.
In 2012 reporting of outcomes of all types of nephrectomy became mandatory in the UK. In all, 148 surgeons in 86 centres prospectively entered data on 6 042 nephrectomies undertaken in 2012. This study is a retrospective analysis of the NSS procedures in the dataset.
A total of 1 044 NSS procedures were recorded and the median (range) surgical volume was 4 (1–39) per consultant and 8 (1–59) per centre. In all, 36 surgeons and 10 centres reported on only one NSS. The indications for NSS were: elective with a tumour of ≤4.5 cm in 59%, elective with a tumour of >4.5 cm in 10%, relative in 7%, imperative in 12%, Von Hippel–Lindau in 1%, and unknown in 11%. The median (range) tumour size was 3.4 (0.8–30) cm. The technique used was minimally invasive surgery in 42%, open in 58%, with conversions in 4%. The histology results were: malignant in 80%, benign in 18%, and unknown in 2%. In patients aged <40 years 36% (36/101) had benign histology vs 17% (151/874) of those aged ≥40 years (P < 0.01). In patients with tumours of <2.5 cm 29% (69/238) had benign histology vs 14% (57/410) with tumours of 2.5–4 cm vs 8% (16/194) with tumours of ≥4 cm (P = 0.02). In patients aged <40 years with of tumours of <2.5 cm 44% (15/34) were benign. The 30-day mortality was 0.1% (1/1 044). There were major complications (Clavien–Dindo grade of ≥IIIa) in 5% (53/1 044). There was an increased risk of complications after extended elective NSS of 19% (19/101) vs elective at 12% (76/621) (relative risk [RR] 1.54; P < 0.01). Margins were recorded in 68% (709/1 044) of the patients, with positive margins identified in 7% (51/709). Positive surgical margins after NSS for pathological T3 (pT3) tumours were found in 47.8% (11/23) vs 6.1% (32/523) for pT1a, tumours (RR 5.61; P < 0.01). In all, 14% (894/6 042) of the patients underwent surgery for T1a tumours: 55% (488/894) by NSS, 42% (377/894) by radical nephrectomy (RN), and in 3% (29/894) the procedure used was unknown. Major complications after occurred in 4.9% (24/488) of NSS vs 1.3% (5/377) of RN (P < 0.01). Limitations included poor reporting of renal function data and no data on tumour complexity.
In its first year, mandatory national reporting has provided several challenging contemporary insights into NSS.
Every Week the Editor-in-Chief selects an Article of the Week from the current issue of BJUI. The abstract is reproduced below and you can click on the button to read the full article, which is freely available to all readers for at least 30 days from the time of this post.
In addition to the article itself, there is an accompanying editorial written by a prominent member of the urological community. This blog is intended to provoke comment and discussion and we invite you to use the comment tools at the bottom of each post to join the conversation.
Finally, the third post under the Article of the Week heading on the homepage will consist of additional material or media. This week we feature a video from Dr. Marios Hadjipavlou, discussing his paper.
If you only have time to read one article this week, it should be this one.
To analyse and compare data from the British Association of Urological Surgeons Nephrectomy Audit for perioperative outcomes of partial (PN) and radical nephrectomy (RN) for T1 renal tumours.
UK consultants were invited to submit data on all patients undergoing nephrectomy between 1 January and 31 December 2012 to a nationally established database using a standard pro forma. Analysis was made on patient demographics, operative technique, and perioperative data/outcome between PN and RN for T1 tumours.
Overall, data from 6 042 nephrectomies were reported of which 1 768 were performed for T1 renal tumours. Of these, 1 082 (61.2%) were RNs and 686 (38.8%) were PNs. The mean age of patients undergoing PN was lower (PN 59 years vs RN 64 years; P < 0.001) and so was the WHO performance score (PN 0.4 vs RN 0.7; P < 0.001). PN for the treatment of T1a tumours (≤4 cm) accounted for 55.6% of procedures, of which 43.9% were performed using a minimally invasive technique. For T1b tumours (4–7 cm), 18.9% of patients underwent PN, in 33.3% of which a minimally invasive technique was adopted. The vast majority of RNs for T1 tumours were performed using a minimally invasive technique (90.3%). Of the laparoscopic PNs, 30.5% were robot-assisted. There was no significant difference in overall intraoperative complications between the RN and PN groups (4% vs 4.3%; P = 0.79). However, PN accounted for a higher overall postoperative complications rate (RN 11.3% vs PN 17.6%; P < 0.001). RN was associated with a markedly reduced risk of severe surgical complications (Clavien Dindo classification grade ≥3) compared with PN even after adjusting for technique (odds ratio 0.30; P = 0.002). Operation time between RN and PN was comparable (141 vs 145 min; P = 0.25). Blood loss was less in the RN group (mean for RN 165 vs PN 323 mL; P < 0.001); however, transfusion rates were similar (3.2% vs 2.6%; P = 0.47). RN was associated with a shorter length of stay (median 4 vs 5 days; P < 0.001). A direct comparison between robot-assisted and laparoscopic PN showed no significant differences in operation time, blood loss, warm ischaemia time, and intraoperative and postoperative complications.
PN was the method of choice for treatment of T1a tumours whereas RN was preferred for T1b tumours. Minimally invasive techniques have been widely adopted for RN but not for PN. Despite the advances in surgical technique, a substantial risk of postoperative complications remains with PN.
In the present issue of BJUI, there is an important study by Hadjipavlou et al. [1], summarizing radical (RN) and partial nephrectomy (PN) practice in the UK in 2012. Specifically, the authors reported the outcomes of ~1 800 patients undergoing either RN or PN for clinical T1 renal masses. Approximately 55% of the patients with cT1a tumours underwent PN, of whom 44% underwent minimally invasive PN. Conversely, in the cohort of patients with cT1b tumours, only ~19% received PN, of whom 33% underwent a minimally invasive procedure. Notably, whereas operating time, transfusion rate and the risk of intraoperative complications was similar for RN and PN, postoperative complications were approximately three times more common in patients who underwent PN, after adjusting for covariates. A sub-analysis comparing robot-assisted and laparoscopic PN failed to show any difference in peri-operative outcomes [1].
The study is important for several reasons. Firstly, it shows a fairly high adoption of PN for cT1a tumours. Although PN is recommended as the standard treatment for small renal masses [2], population-based studies have shown that there has been limited adoption of PN outside referral centres [3, 4], especially in the USA. Conversely, the present data from UK show more encouraging results, maximizing the benefit of nephron preservation; however, although PN might be more challenging in cT1b tumours and the available evidence in favour of PN in such a setting is less compelling, the adoption of PN was lower in such tumours. Efforts should be made to popularize such an approach whenever feasible.
Secondly, the study showed that a minority of the PN procedures were performed with a minimally invasive approach. Although we can agree that nephron preservation is more important than a minimally invasive approach in the long term for most patients, an increasing number of publications and growing clinical experience suggest that laparoscopic, and, above all, robot-assisted PN could represent the ideal solution. Although the number of minimally invasive PNs should increase with increased diffusion of DaVinci platforms, major efforts should be made to expand the number of patients in whom the morbidity of the traditional open PN approach can be avoided. In this context, regionalization of care for PN, as for other major oncological procedures, could be an excellent solution.
Thirdly, the significant rise in the risk of postoperative complications observed after PN could allow better selection of patients to undergo either PN or RN. For example, where surgery is indicated, frail comorbid patients, in whom the risk of perioperative complications should be minimized and who would benefit less from nephron preservation, could be better treated by laparoscopic RN or, probably, robot-assisted PN as performed by very experienced surgeons.
Finally, the study failed to show major differences between laparoscopic and robot-assisted PN. Although this finding is in line with data from systematic reviews of the literature [5], the present data from a large cohort of surgeons are more solid. The lack of data on patient selection, previous laparoscopic and robot-assisted surgery, annual surgical volume and tumour characteristics according to nephrometry scores, however, does not allow us to draw definitive conclusions on the issue. In our opinion, robot-assisted surgery might offer major significant benefits during PN in terms of quicker and more accurate tumour dissection, improved renorrhaphy with consequent shorter ischaemia time, lower risk of complications and a shorter learning curve as compared with pure laparoscopic PN.
Unfortunately, no analysis stratified by centre and/or surgeon volume was provided in the present paper. As with other major surgical procedures, some studies suggest that case volume may have a major impact on outcome [6]. It would have been interesting to see such a relationship analysed in the present cohort involving almost 300 surgeons from more than 100 institutions. Despite the large number of cases analysed, however, it is likely that these data depict the outcomes of RN and PN in a low-volume setting (an average of approximately six cases per year in total).
Finally, alternative approaches such as percutaneous or laparoscopic cryoablation are gaining popularity for the treatment of small renal masses in selected cases [2]. Although long-term oncological outcomes of such procedures are lacking, the available evidence suggests good short-term efficacy and safety for cryoablation in patients with small renal masses. The presence of data on such treatments to compare with the surgery results reported in the present cohort would also have been of interest.
To analyse and compare data from the British Association of Urological Surgeons Nephrectomy Audit for perioperative outcomes of partial (PN) and radical nephrectomy (RN) for T1 renal tumours.
UK consultants were invited to submit data on all patients undergoing nephrectomy between 1 January and 31 December 2012 to a nationally established database using a standard pro forma. Analysis was made on patient demographics, operative technique, and perioperative data/outcome between PN and RN for T1 tumours.
Overall, data from 6 042 nephrectomies were reported of which 1 768 were performed for T1 renal tumours. Of these, 1 082 (61.2%) were RNs and 686 (38.8%) were PNs. The mean age of patients undergoing PN was lower (PN 59 years vs RN 64 years; P < 0.001) and so was the WHO performance score (PN 0.4 vs RN 0.7; P < 0.001). PN for the treatment of T1a tumours (≤4 cm) accounted for 55.6% of procedures, of which 43.9% were performed using a minimally invasive technique. For T1b tumours (4–7 cm), 18.9% of patients underwent PN, in 33.3% of which a minimally invasive technique was adopted. The vast majority of RNs for T1 tumours were performed using a minimally invasive technique (90.3%). Of the laparoscopic PNs, 30.5% were robot-assisted. There was no significant difference in overall intraoperative complications between the RN and PN groups (4% vs 4.3%; P = 0.79). However, PN accounted for a higher overall postoperative complications rate (RN 11.3% vs PN 17.6%; P < 0.001). RN was associated with a markedly reduced risk of severe surgical complications (Clavien Dindo classification grade ≥3) compared with PN even after adjusting for technique (odds ratio 0.30; P = 0.002). Operation time between RN and PN was comparable (141 vs 145 min; P = 0.25). Blood loss was less in the RN group (mean for RN 165 vs PN 323 mL; P < 0.001); however, transfusion rates were similar (3.2% vs 2.6%; P = 0.47). RN was associated with a shorter length of stay (median 4 vs 5 days; P < 0.001). A direct comparison between robot-assisted and laparoscopic PN showed no significant differences in operation time, blood loss, warm ischaemia time, and intraoperative and postoperative complications.
PN was the method of choice for treatment of T1a tumours whereas RN was preferred for T1b tumours. Minimally invasive techniques have been widely adopted for RN but not for PN. Despite the advances in surgical technique, a substantial risk of postoperative complications remains with PN.